Re: [irsg] Resending: Page numbers in RFCs questions / preferences

[Benjamin Kaduk <[e-mail]>]

Hi Jeff,

On Tue, Oct 27, 2020 at 03:38:08PM -0400, Jeffrey Haas wrote:
> Ben,
> 
> On Mon, Oct 26, 2020 at 02:51:17PM -0700, Benjamin Kaduk wrote:
> > > The input XML already provides blocks.  It should be fairly reasonable for
> > > the tool to provide you something like "this is section X.Y, ¶5".
> > 
> > Like https://www.rfc-editor.org/rfc/rfc8815.html#section-3.1-5 ?
> 
> You seem to have missed the paragraph 5 symbol I added.
> 
> I regularly use the html anchor notation above. :-)

I am not sure which of us is confused -- the link I provided is directly to
the fifth paragraph of the section (the "-5" suffix effectuates that
functionality).  So, if I assume that you knew that, I'm left with the
conclusion that the "paragraph" part of the reference/link needs to be more
prominent to someone that does not understand the URL fragment's structure.
Which is perhaps a valid position to hold, but I'm not entirely confident
if that is the position you are holding.

> If it wasn't for the fact that our XML is of highly variable quality

Indeed.  And there does not seem to be much appetite for fixing that -- I
had asked about even something as simple as consistent indentation for
nested XML elements and got a response along the lines of "if authors care,
they can put in the work to make the XML look pretty".  Which is
unsatisfying to me, but I apparently failed to make a persuasive case for
it.

-Ben

> I'd simply note that XPath provides sufficient addressability.  But it's
> less easy to interact with it in plain text emails and still awkward in
> richer formats.
> 
> /section[@title='Security Considerations']/t[5]  as an example of the fifth
> text block of the security considerations section.
> 
> 
> -- Jeff